为了改善数字全息重构像质量以达到可实际利用的目的，首先采用以马赫－曾德尔干涉仪为基础的三维物体全息实验光路，运用单次曝光的方法，记录了真实三维物体的数字全息图。由于对该全息图直接计算得到的重构像，受零级衍射光斑和散斑噪声的影响较大，使得重构像质量难以令人满意。为此采用了一套独特的多重滤波数字图像处理方法，即综合运用小波分析与收缩加权平均滤波对实验获得的数字全息图及其数字重构像进行滤波处理，成功地消除了数字重构像中的零级衍射光斑、减小了散斑噪声的影响，得到逼真的三维物体重构像。实验结果表明，该方法简便、实用，可显著提高三维物体重构像的质量。由于采用单次曝光，无需记录多幅图像，此方法便于在实时图像分析处理等领域中应用。

For improving the quality of the reconstructed image calculated from the hologram,a 3-D object digital hologram was successfully recorded by using single exposure method and 3-D object holographic optical setup based on Mach-Zehnder interferometer. The quality of the reconstructed image directly calculated from the recorded hologram was unsatisfactory because of the presence of the zero-order diffraction and the speckle noise. A set of digital image processing algorithms was then proposed to process the digital hologram and the reconstructed image with multi-filtering technique. The method synthetizing the wavelet filtering and the contracted weighted mean filtering is capable to eliminate the zero-order diffraction component and reduce the speckle noise to obtain realistic reconstructed image. The experimental result shows that this simple method is very effective for improving the quality of the reconstructed image. For recording hologram only once,it is convenient for application in the real time image processing fields.